Performance Metric Recommendations for Handling Multi-Party Electronic Communications

1. A computer-implemented method for processing one or more performance metric recommendations for an agent profile, the computer-implemented method comprising: generating, via one or more processors, an agent feature data object for the agent profile representing an agent based at least in part on a plurality of communication data objects representing a plurality of communications associated with the agent profile; processing, via the one or more processors, the agent feature data object using an agent group identifier machine learning model to generate an agent group data object for the agent profile; identifying, via the one or more processors, a top agent performer data object based at least in part on the agent group data object generated for the agent profile; generating, via the one or more processors, an agent assessment data object for the agent profile, wherein the agent assessment data object represents a performance evaluation for the agent represented by the agent profile in relation to one or more of the plurality of communications; processing, via the one or more processors, the agent assessment data object and the top agent performer data object using a comparison machine learning model to generate one or more inferred performance gap data objects; generating, via the one or more processors, the one or more performance metric recommendations based at least in part on the one or more inferred performance gap data objects; and performing, via the one or more processors, one or more performance-related actions based at least in part on the one or more performance metric recommendations.

2. The computer-implemented method of claim 1 , wherein generating the agent feature data object for the agent profile comprises: generating one or more text transcripts from the plurality of communication data objects; and processing the one or more text transcripts using a transcription processing machine learning model to generate a transcription-based embedding data object for the agent feature data object.

3. The computer-implemented method of claim 2 , wherein the transcription processing machine learning model comprises bidirectional encoder representations from a transformer-based natural language processing machine learning model.

4. The computer-implemented method of claim 2 , wherein generating the agent feature data object for the agent profile further comprises: processing one or more behavioral detection data objects for the agent profile using a behavioral processing machine learning model to generate a behavioral-based embedding data object, wherein the one or more behavioral detection data objects identify behavioral attributes recorded in relation to an observed behavior of the agent represented by the agent profile during the plurality of communications; and processing the transcription-based embedding data object and the behavioral-based embedding data object using an agent feature merger machine learning model to generate the agent feature data object.

5. The computer-implemented method of claim 1 further comprising: generating a party feature data object for a party profile representing a party; and processing, via the one or more processors, the party feature data object using a talking point identifier machine learning model to generate a plurality of recommended talking point data objects for the party profile.

6. The computer-implemented method of claim 5 , wherein the one or more performance-related actions comprise causing presentation of a personalized simulation training characterized by a simulated communication data object between the agent represented by the agent profile and the party represented by the party profile, and wherein the simulated communication data object is determined based at least in part on the plurality of recommended talking point data objects.

7. The computer-implemented method of claim 5 , wherein generating the party feature data object for the party comprises: generating one or more text transcripts from a plurality of communication data objects associated with the party profile; and processing, via the one or more processors, the one or more text transcripts using the transcription processing machine learning model to generate a transcription-based embedding data object of the party feature data object.

8. The computer-implemented method of claim 7 , wherein generating the party feature data object for the party further comprises: processing, via the one or more processors, one or more party feature data objects associated with the party profile using a party feature processing machine learning model to generate a party-based embedding data object; and processing, via the one or more processors, the transcription-based embedding data object for the party profile and the party-based embedding data object using a party feature merger machine learning model to generate the party feature data object.

9. An apparatus for processing one or more performance metric recommendations for an agent profile, the apparatus comprising at least one processor and at least one memory including program code, the at least one memory and the program code configured to, with the at least one processor, cause the apparatus to at least: generate an agent feature data object for the agent profile representing an agent based at least in part on a plurality of communication data objects representing a plurality of communications associated with the agent profile; process the agent feature data object using an agent group identifier machine learning model to generate an agent group data object for the agent profile; identify a top agent performer data object based at least in part on the agent group data object generated for the agent profile; generate an agent assessment data object for the agent profile, wherein the agent assessment data object represents a performance evaluation for the agent represented by the agent profile in relation to one or more of the plurality of communications; process the agent assessment data object and the top agent performer data object using a comparison machine learning model to generate one or more inferred performance gap data objects; generate the one or more performance metric recommendations based at least in part on the one or more inferred performance gap data objects; and have one or more performance-related actions performed based at least in part on the one or more performance metric recommendations.

10. The apparatus of claim 9 , wherein the at least one memory and the program code are configured to, with the at least one processor, cause the apparatus to generate the agent feature data object for the agent profile by: generating one or more text transcripts from the plurality of communication data objects; and processing the one or more text transcripts using a transcription processing machine learning model to generate a transcription-based embedding data object for the agent feature data object.

11. The apparatus of claim 10 , wherein the at least one memory and the program code are configured to, with the at least one processor, cause the apparatus to generate the agent feature data object for the agent profile by: processing one or more behavioral detection data objects for the agent profile using a behavioral processing machine learning model to generate a behavioral-based embedding data object, wherein the one or more behavioral detection data objects identify behavioral attributes recorded in relation to an observed behavior of the agent represented by the agent profile during the plurality of communications; and processing the transcription-based embedding data object and the behavioral-based embedding data object using an agent feature merger machine learning model to generate the agent feature data object.

12. The apparatus of claim 9 , wherein the at least one memory and the program code are configured to, with the at least one processor, cause the apparatus to: generate a party feature data object for a party profile representing a party; and process the party feature data object using a talking point identifier machine learning model to generate a plurality of recommended talking point data objects for the party profile.

13. The apparatus of claim 12 , wherein the one or more performance-related actions comprise causing presentation of a personalized simulation training characterized by a simulated communication data object between the agent represented by the agent profile and the party represented by the party based at least in part on the plurality of recommended talking point data objects.

14. The apparatus of claim 12 , wherein the at least one memory and the program code are configured to, with the at least one processor, cause the apparatus to generate the party feature data object for the party by: generating one or more text transcripts from a plurality of communication data objects associated with the party profile; and processing the one or more text transcripts using the transcription processing machine learning model to generate a transcription-based embedding data object of the party feature data object.

15. The apparatus of claim 14 , wherein the at least one memory and the program code are configured to, with the at least one processor, cause the apparatus to generate the party feature data object for the party by: processing one or more party feature data objects associated with the party profile using a party feature processing machine learning model to generate a party-based embedding data object; and processing the transcription-based embedding data object for the party profile and the party-based embedding data object using a party feature merger machine learning model to generate the party feature data object.

16. A non-transitory computer storage medium comprising instructions for processing one or more performance metric recommendations for an agent profile, the instructions being configured to cause one or more computer processors to at least perform operations configured to: generate an agent feature data object for the agent profile representing an agent based at least in part on a plurality of communication data objects representing a plurality of communications associated with the agent profile; process the agent feature data object using an agent group identifier machine learning model to generate an agent group data object for the agent profile; identify a top agent performer data object based at least in part on the agent group data object generated for the agent profile; generate an agent assessment data object for the agent profile, wherein the agent assessment data object represents a performance evaluation for the agent represented by the agent profile in relation to one or more of the plurality of communications; process the agent assessment data object and the top agent performer data object using a comparison machine learning model to generate one or more inferred performance gap data objects; generate the one or more performance metric recommendations based at least in part on the one or more inferred performance gap data objects; and have one or more performance-related actions performed based at least in part on the one or more performance metric recommendations.

17. The non-transitory computer storage medium of claim 16 , wherein the instructions are configured to cause the one or more computer processors to at least perform operations configured to generate the agent feature data object for the agent profile by: generating one or more text transcripts from the plurality of communication data objects; and processing the one or more text transcripts using a transcription processing machine learning model to generate a transcription-based embedding data object for the agent feature data object.

18. The non-transitory computer storage medium of claim 17 , wherein the instructions are configured to cause the one or more computer processors to at least perform operations configured to generate the agent feature data object for the agent profile by: processing one or more behavioral detection data objects associated with the agent profile using a behavioral processing machine learning model to generate a behavioral-based embedding data object, wherein the one or more behavioral detection data objects identify behavioral attributes recorded in relation to an observed behavior of the agent represented by the agent profile during the plurality of communications; and processing the transcription-based embedding data object and the behavioral-based embedding data object using an agent feature merger machine learning model to generate the agent feature data object.

19. The non-transitory computer storage medium of claim 16 , wherein the instructions are configured to cause the one or more computer processors to at least perform operations configured to: generate a party feature data object for a party profile; and process the party feature data object using a talking point identifier machine learning model to generate a plurality of recommended talking point data objects for the party profile.

20. The non-transitory computer storage medium of claim 19 , wherein the one or more performance-related actions comprise causing presentation of a personalized simulation training characterized by a simulated communication data object between the agent represented by the agent profile and the party represented by the party profile, and wherein the simulated communication data object is determined based at least in part on the plurality of recommended talking point data objects.

21. The non-transitory computer storage medium of claim 19 wherein the instructions are configured to cause the one or more computer processors to at least perform operations configured to generate the party feature data object for the party by: generating one or more text transcripts from a plurality of communication data objects associated with the party profile; processing the one or more text transcripts using the transcription processing machine learning model to generate a transcription-based embedding data object of the party feature data object; processing one or more party feature data objects associated with the party profile using a party feature processing machine learning model to generate a party-based embedding data object; and processing the transcription-based embedding data object for the party profile and the party-based embedding data object using a party feature merger machine learning model to generate the party feature data object.